Filtering databases and chemical libraries
Filtering databases and chemical libraries
复制标题
DOI：
10.1023/a:1020829519597
复制
发表时间：
2002-05-01
影响因子：
3.5
通讯作者：
Walters, WP
中科院分区：
文献类型：
--
作者：
Charifson, PS;Walters, WP
在当前高通量化学和筛选的环境下，有许多化合物可以合成和筛选。然而，最近的经验(和常识)表明，并不是所有可以合成的化合物或存在于化合物集合中的所有化合物都值得在任何给定的目标上进行筛选。在药物筛选库上执行HTS通常不到1%[1]。虽然有许多策略可以尝试提高“命中率”(即浓缩)，包括多样性方法和有重点的文库设计，但实现这一目标的最简单方法是删除在药物发现项目的“先导生成”阶段提供有用信息的概率较低的化合物。这种简化论的方法利用了最基本的科学原理之一：清除过程。公平地说，这种过滤方法虽然有用，但也有很大的局限性，最好与其他技术结合使用。例如，通常的做法是除去所有具有不希望看到的官能团的潜在化合物，然后对其余化合物进行多样性分析。另一个例子是将所有符合定义的化学成分且ClogP小于某个期望值的化合物对接到蛋白质结合部位。然后，人们可以只以图形方式评估那些接触得分“良好”的化合物，并选择一组最终的化合物进行合成和/或筛选。在每一种情况下，都有一定程度的主观性被用来定义什么是“好的”或“不想要的”。这些“门槛”值通常来自给定组织内的经验或整个制药业的集体经验。与如何使用这种过滤器有关的另一个关键问题是何时使用这种过滤器。通常，使用本章中讨论的过滤器类型
In the current climate of high-throughput chemistry and screening, there are many compounds that can be synthesized and screened. Recent experience (and common sense) suggests, however, that not all compounds which can be synthesized or which are present in compound collections are worthy of screening on any given target.‘Hit-rates’ are typically less than 1%[1] for HTS performed on pharmaceutical screening libraries. Although many strategies exist for attempting to improve ‘hit-rates’(ie enrichment) including diversity approaches and focused library design, the simplest way of accomplishing this goal is to remove compounds which have a low probability of providing useful information at the ‘lead-generation’stage of a drug discovery project. This reductionist approach utilizes one of the most fundamental scientific tenets: the process of elimination.It would be fair to state that such filtration approaches, although useful, possesses significant limitations and are best used in conjunction with other techniques. For example, it is a common practice to remove all potential compounds possessing ‘undesirable’functional groups and then perform a diversity analysis on the remaining compounds. Another example would be to take all compounds consistent with a defined chemistry with a ClogP less than some desired value and dock them into a protein binding site. One might then graphically evaluate only those compounds with ‘favorable’contact scores and select a final set of compounds for synthesis and/or screening. In each of these cases, there is a degree of subjectivity employed in defining what is ‘favorable’or ‘undesirable’. These ‘threshold’values are usually derived from experiences within a given organization or the collective experience across the pharmaceutical industry. Another key issue related to how such filters are employed is when to use such filters. Typically, the types of filters discussed in this chapter are employed